Albeit being two separate diagnoses, schizophrenia (SZ) and bipolar disorder (BP) share psychotic symptoms, chronic courses and cognitive deficits, as well as genetic determinants from clinical, epidemiological and genetic findings. The underlying mechanism(s) accounting for such a relation is far from clear. This project is designed to improve the understanding of pathways by which genetic predisposition influences brain structure abnormalities associated with clinical symptoms, through combining the advances of genetic and neuroimaging techniques. We propose a hierarchical study that will leverage a large-sample genomic analysis, and specific genetic neuroimaging association analyses. First, we will use a very large sample of single nucleotide polymorphisms and copy number variations from the Psychiatric Genomics Consortium (PGC) SZ and BP patients in order to reliably identify the candidate chromosome regions. Within the candidate regions we will apply multivariate imaging-genetic methods to extract specific genetic factors directly associated with brain structural networks that differentiate patients from controls, using a locally collected sample. Gray matter concentration and white matter integrity within brain networks will be analyzed for their associations with both genetic profiles and patient diagnoses. Through comparison of such associations in SZ, BP and healthy control groups, we will be able to identify common and unique neuropathological mechanisms underlying the two disorders. Such genetic and neural features can act as effective biomarkers for refined patient categorization.